In his Nature article, Spiegelman wrote that exercise spurs the production of a protein called Fndc5, which then breaks apart. One of the pieces, a chemical messenger secreted into the blood, activated the conversion of white fat into brown fat. Spiegelman called it irisin, after the Greek messenger goddess Iris. Mice given Fndc5 lost weight and had improved blood sugar levels.

Mainstream media publications took note. “Remarkable new insights into how exercise affects the body,” wrote the New York Times. “Lucky Mice: Researchers have Found a Protein that Mimics the Effects of Exercise,” proclaimed Time magazine. Bloomberg News weighed in as well with a story headlined “Exercise-Related Hormone May Help Obesity.”

Boston-based Third Rock Ventures, which specializes in creating early-stage companies in cutting-edge areas of science and health care, jumped in. Third Rock, which has $1.3 billion under management, set up Ember Therapeutics in December 2011, committing $34 million, and bought the rights for irisin and other molecules.

Spiegelman came on as scientific founder, and Chief Executive Officer Lou Tartaglia said Ember hoped to test a version of irisin in humans as soon as two years from the company’s launch. Third Rock’s investment was not unusual in size; in the same year, it launched Sage Therapeutics Inc., a biotech tackling rare central nervous system disorders, with $35 million in funding and led the $40 million fundraising to create cancer drugmaker Blueprint Medicines Corp.

Third Rock Backs Out

Both Sage and Blueprint Medicines have successfully gone public -- Sage’s shares have more than tripled since its July 2014 IPO, and Blueprint’s shares have gained 62 percent since going public.

Yet Ember wasn’t moving forward at the same pace. Third Rock, which has declined to say how much of its initial investment was spent prior to closing Ember, announced in March that it had sold some of Ember’s intellectual property to closely held Mariel Therapeutics Inc. Mariel wasn’t interested in the rights for irisin, which have been returned to Harvard.

Ember ceased lab operations. “It would take a longer time horizon than anticipated to sort through the complex biology and pathways of brown fat,” explained Third Rock partner Kevin Starr.

That same month, Duke professor Harold Erickson wrote an article criticizing Spiegelman’s and other academics’ irisin research. The testing kits used to measure irisin in more than 100 studies were flawed, he said in the article, which was published in the journal Scientific Reports. Comparing two tests, a commonly used one called an Elisa and another called a Western Blot, Erickson and a team of researchers found that when the Elisa said it had found irisin, the Western Blot said those proteins were too big to be irisin.

‘No Credible Evidence’